Fermilab Antiproton Source, Recycler Ring, and

Main Injector

Sergei Nagaitsev, Fermilab, Batavia, IL 60510

Introduction

The antiproton source for a proton-antiproton collider at Fermilab was

proposed in 1976 [1]. The proposal argued that the requisite luminosity (~1029

cm-2

sec-1

) could be achieved with a facility that would produce and cool

approximately 1011

antiprotons per day. Funding for the Tevatron I project (to

construct the Antiproton source) was initiated in 1981 and the Tevatron ring itself

was completed, as a fixed target accelerator, in the summer of 1983 and the

Antiproton Source was completed in 1985. At the end of its operations in 2011,

the Fermilab antiproton production complex consisted of a sophisticated target

system, three 8-GeV storage rings (namely the Debuncher, Accumulator and

Recycler), 25 independent multi-GHz stochastic cooling systems, the world’s

only relativistic electron cooling system and a team of technical experts equal to

none. Sustained accumulation of antiprotons was possible at the rate of greater

than 2.5×1011

per hour. Record-size stacks of antiprotons in excess of 3×1012

were accumulated in the Accumulator ring and 6×1012

in the Recycler. In some

special cases, the antiprotons were stored in rings for more than 50 days. Note,

that over the years, some 1016

antiprotons were produced and accumulated at

Fermilab, which is about 17 nanograms and more than 90% of the world’s total

man-made quantity of nuclear antimatter.

The accelerator complex at Fermilab supported a broad physics program

including the Tevatron Collider Run II [2], neutrino experiments using 8 GeV

and 120 GeV proton beams, as well as a test beam facility and other fixed target

experiments using 120 GeV primary proton beams. The following sections

provide a brief description of Fermilab accelerators as they operated at the end of

the Collider Run II (2011).

FERMILAB-FN-0957-AD

Operated by Fermi Research Alliance, LLC under Contract No. De-AC02-07CH11359 with the United States Department of Energy.

2

1. The Proton Source

The Proton Source consists of the Pre-Accelerator (Pre-Acc), the Linac, and

the Booster. For operational redundancy, there were two independent 750-kV

Pre-Acc systems which provided H- ions for acceleration through the Linac.

Each Pre-Acc was a Cockcroft-Walton accelerator having its own magnetron-

type H- source running at a 15 Hz repetition rate, a voltage multiplier to generate

the 750 kV accelerating voltage, and a chopper to set the beam pulse length

going into the Linac. The typical H- source output current is 40-60 mA.

The Linac accelerated H- ions from 750 keV to 400 MeV. Originally, the

Linac was a 200 MeV machine made entirely of Alvarez-style drift tube tanks

[3], but a 1991 upgrade replaced four of the drift tubes with side coupled cavities

to allow acceleration up to 400 MeV [4]. Today, the low energy section (up to

116 MeV) is made of drift tube tanks operating with 201.25 MHz RF fed from

triod-based 5 MW power amplifier tubes. The high energy section (116 – 400

MeV) consists of 7 side-coupled cavity girders powered by 805-MHz, 12-MW

Klystrons providing a gradient of ≈ 7 MV/m. A transition section between the

two linac sections provides the optics matching and rebunching into the higher

frequency RF system. The nominal beam current in the Linac is 30-35 mA.

The Booster is a 474 meter circumference, rapid-cycling synchrotron

ramping from 400 MeV to 8 GeV at a 15 Hz repetition rate. (Note that while the

magnets ramp at 15 Hz, beam is not present on every cycle.) Multi-turn injection

is achieved by passing the incoming H- ions through 1.5 m thick (300 g/cm

2)

carbon stripping foils as they merge with the circulating proton beam on a

common orbit. The 96 10-foot long combined-function Booster gradient

magnets are grouped into 24 identical periods in a FOFDOOD lattice [5]. The

Booster RF system (harmonic number = 84) consists of 19 cavities (18

operational + 1 spare) that must sweep from 37.9-52.8 MHz as the beam velocity

increases during acceleration. The ferrite tuners and power amplifiers are

mounted on the cavities in the tunnel. The cavities provide a total of ≈750 kV

per turn for acceleration. The Booster transition energy (4.2 GeV) occurs at 17

ms in the cycle. The Booster throughput efficiency is 85-90% for typical beam

intensities of 4.5-5.0×1012

protons per pulse. The majority of the proton flux

through the Booster is delivered to the 8 GeV and 120 GeV neutrino production

targets.

3

2. The Main Injector

The Main Injector [6] (MI) is a 3319.4-m circumference synchrotron, which

can accelerate both proton and antiproton beams from 8 GeV up to 150 GeV. It

has a FODO lattice using conventional, separated function dipole and quadrupole

magnets. There are also trim dipole and quadrupoles, skew quadrupole, sextupole

and octupole magnets in the lattice. Since the Main Injector circumference is

seven times the Booster circumference, beam from multiple consecutive Booster

cycles, called batches, can be injected around the Main Injector. In addition, even

higher beam intensity can be accelerated by injecting more than seven Booster

batches through the process of slip-stacking: capturing one set of injected proton

batches with one RF system, decelerating them slightly, then capturing another

set of proton injections with another independent RF system, and merging them

prior to acceleration. There are 18 53-MHz RF cavities, grouped into 2

independently controlled systems to allow slip-stacking and the flexibility for

maintenance. Beam-loading compensation and active damping systems have

been implemented to help maintain beam stability. For beam injections into the

Tevatron, coalescing of several 53 MHz bunches of protons and antiprotons into

single, high intensity bunches also requires a 2.5 MHz system for bunch rotations

and a 106 MHz cavity to flatten the potential when recapturing beam into the

single 53 MHz bunch to be injected into the Tevatron ring. A set of collimators

was installed in the Main Injector to help localize beam losses to reduce

widespread activation of ring components in the tunnel.

The Main Injector supports various operational modes for delivering beam

across the complex. For the antiproton and neutrino production, up to 11 proton

batches from the Booster were injected and slip-stacked prior to acceleration.

After reaching 120 GeV, 2 batches were extracted to the antiproton production

target while the remaining 9 batches were extracted to the NuMI neutrino

production target (Figure 1). At its peak performance, the Main Injector can

sustain 400 kW delivery of 120 GeV proton beam power at 2.2 sec cycle times.

The Main Injector also provides 120 GeV protons in a 4 sec long slow-spill

extracted to the Switchyard as a primary beam or for production of secondary

and tertiary beams for the Meson Test Beam Facility and other fixed-target

experiments. In addition, the Main Injector served as an “effective” transport

line for 8 GeV antiprotons being transferred from the Accumulator to the

Recycler for later use in the Tevatron. Protons from the Booster and antiprotons

from the Recycler were accelerated to 150 GeV in the Main Injector and

4

coalesced into higher intensity bunches for injection into the Tevatron for a

colliding beam store.

Figure 1: The Main Injector cycle illustrating an 11-batch proton injection,

acceleration and extraction.

3. The Antiproton Source

The Antiproton Source [7] had 3 main parts: the Target Station, the

Debuncher, and the Accumulator. Each of these is described briefly below while

outlining the steps of an antiproton production cycle. In the Target Station,

batches of 120-GeV protons (~8×1012

per batch), delivered from the Main

Injector, strike the Inconel (a nickel-iron alloy) target every 2.2 sec. The beam

spot on the target can be controlled by a set of quadrupole magnets. The target is

rotated between beam pulses to spread the depletion and damage uniformly

around its circumference. The shower of secondary particles, emanating from

the target, was focused both horizontally and vertically by a pulsed, high current

lithium lens that can provide up 1000 T/m gradient.

5

Figure 2: The Antiproton Source consists of the Debuncher ring (outer

circumference, left) and the Accumulator ring (inner circumference, right).

Downstream of the Li lens was a pulsed dipole magnet, which steered

negatively-charged particles at 8.9-GeV/c momentum into the transport line

toward the Debuncher. A collimator between the lens and pulsed magnet was

installed to help protect the pulsed magnet from radiation damage as the

incoming primary proton beam intensity increased with proton slip-stacking in

Main Injector.

The Debuncher and Accumulator (Figure 2) were both triangular-shaped

rings of conventional magnets sharing the same tunnel. While the Debuncher

had a FODO lattice, the Accumulator lattice had particular features needed for

cooling and accumulating antiprotons with stochastic cooling systems. A total of

21 independent stochastic cooling systems were implemented in the Accumulator

and Debuncher [8]. Such a variety of cooling systems was possible after a series

of development efforts [9, 10] allowing for more robust and less expensive pick-

up arrays.

The ~2×108 bunches of antiprotons entering the Debuncher from the

transport line retained the 53-MHz bunch structure from the primary proton beam

on the target. A 53-MHz RF system (harmonic number = 90) was used for the

bunch rotation and debunching of the antiprotons into a continuous beam with a

low momentum spread. An independent 2.4 MHz RF system provided a barrier

bucket to allow a gap for extraction to the Accumulator. Stochastic cooling

6

systems reduced the transverse emittance from 300 to 30 π mm-mrad (rms,

normalized) and momentum spread from 0.30% to <0.14% prior to injection into

the Accumulator.

Figure 3: The frequency (energy) distribution of antiprotons in the

Accumulator highlighting incoming antiprotons (left), the stacktail beam

(middle) being cooled and decelerated toward the core (right). Higher beam

energy is to the left, lower energy is to the right.

In the Accumulator, antiprotons were momentum-stacked and cooled by a

series of RF manipulations and stochastic cooling. The incoming antiprotons

were captured and decelerated by 60 MeV with a 53-MHz RF system (harmonic

number = 84) to the central orbit where the beam was adiabatically debunched.

Before the next pulse of antiprotons arrived (every 2.2 sec), the so-called

stacktail momentum stochastic cooling system [11] decelerated the antiprotons

another 150 MeV toward the core orbit where another set of independent betatron

and momentum stochastic cooling systems provided additional cooling while

building a “stack” of antiprotons. Figure 3 illustrates the frequency (energy)

distribution of antiprotons in the Accumulator. Figure 4 shows the average

antiproton accumulation rates since 1994; typical values at the end of Run II

were in the range 24-26 ×1010

/hr, with a maximum recorder rate of 28.5 ×1010

/hr.

7

Figure 4: The average antiproton accumulation rate since 1994 and during

all of Collider Run II (including the production in the Antiproton Source and

storage in the Recycler). ). Note that some data points at the highest rates, in

particular in the early years, are merely artifacts of the data acquisition and

logging system.

3. The Recycler

The Recycler [12] is a permanent-magnet, fixed momentum (8.9-GeV/c)

storage ring located in the Main Injector tunnel (Figure 5). As conceived the

Recycler would provide storage for very large numbers of antiprotons (up to

6×1012

) and would increase the effective production rate by recapturing unused

antiprotons at the end of collider stores (hence the name Recycler). Recycling of

antiprotons was determined to be ineffective and was never implemented.

However, the Recycler was used as a final antiproton cooling and storage ring for

accumulating significantly larger stashes (so called to differentiate from

Accumulator ‘stacks’) of antiprotons than can be accumulated in the Antiproton

Accumulator. The main Recycler magnets are combined-function strontium

ferrite permanent magnets arranged in a FODO lattice. Trim electro-magnets are

used to provide orbit and lattice corrections.

8

Figure 5: The Recycler (top) and the Main Injector (bottom) rings installed

in a common tunnel.

An important feature of the Recycler was an electron cooling system [13]. It

augmented the Recycler’s cooling capability and complemented the stochastic

cooling system and its inherent limitations. The Pelletron (an electrostatic

accelerator manufactured by the National Electrostatics Corp.) provided a 4.3

MeV electron beam (up to 500 mA) which overlaped with the 8-GeV

antiprotons, circulating in the Recycler, in a 20-m long section and cooled the

antiprotons both transversely and longitudinally. Figure 6 shows the schematic

layout of the Fermilab electron cooling system. The dc electron beam was

generated by a thermionic gun, located in the high-voltage terminal of the

electrostatic accelerator. This accelerator was incapable of sustaining dc beam

currents to ground in excess of about 100 µA. Hence, to attain the electron dc

current of 500 mA, a recirculation scheme was employed. A typical relative

beam current loss in such a scheme was 2×10-5

.

9

Figure 6: Schematic layout of the Recycler electron cooling system and

accelerator cross-section (inset).

The Fermilab system employed a unique beam transport scheme [14]. The

electron gun was immersed in a solenoidal magnetic field, which created a beam

with large angular momentum. After the beam was extracted from the magnetic

field and accelerated to 4.3 MeV, it was transported to the 20-m long cooling

section solenoid using conventional focusing elements (as opposed to low-energy

electron coolers where the beam remains immersed in a strong magnetic field at

all times). The cooling section solenoid removed this angular momentum and the

beam was made round and parallel such that the beam radius, a, produced the

same magnetic flux, Ba2, as at the cathode. The magnetic field in the cooling

section was quite weak (100 G), therefore the kinetics of the electron-antiproton

scattering was weakly affected by the magnetic field.

After becoming operational in September 2005, electron cooling in the

Recycler allowed for significant improvements in Tevatron luminosity by

providing higher intensity antiprotons with smaller emittances. With electron

cooling, the Recycler has been able to store up to 6×1012

antiprotons. In routine

operations, the Recycler accumulated 3.5-4.0×1012

antiprotons with a ~200-hr

lifetime for injection to the Tevatron [15].

Among other unique features of the Recycler was the so-called barrier-

bucket rf system [16] which allowed for crucial longitudinal beam manipulations

of the accumulated antiproton beam.

10

5. Antiproton Flow and Tevatron Shot-Setup

As mentioned previously, stacks of freshly produced antiprotons were stored

temporarily in the Accumulator. The Accumulator antiproton stack was

periodically transferred to the Recycler where electron cooling allowed for a

much larger antiproton intensity to be accumulated with smaller emittances.

Typically 22-25×1010

antiprotons were transferred to the Recycler every ~60

minutes. Prior to electron cooling in the Recycler, antiprotons destined for the

Tevatron were extracted from the Accumulator only. Since late 2005, all

Tevatron antiprotons were extracted from the Recycler only. Figure 7 illustrates

the flow of antiprotons between the Accumulator, Recycler and Tevatron over a

1 week period.

Figure 7: Production and transfers of antiprotons between the Accumulator

and Recycler over 1 week of operation. While the Tevatron has a colliding beam

store, small stacks of antiprotons are produced and stored in the Accumulator,

and then periodically transferred to the Recycler in preparation for the

subsequent Tevatron fill.

A typical Tevatron collider fill cycle is shown in Figure 8 [17]. First, proton

bunches were injected two at a time on the central orbit. Then, electrostatic

separators were powered to put the protons onto a helical orbit. Antiproton

bunches were then injected (four bunches at a time) into gaps between the three

proton bunch trains. After each group of 3 antiproton transfers, the gaps were

11

cleared for the subsequent set of transfers by “cogging” the antiprotons –

changing the antiproton RF cavity frequency to let them slip longitudinally

relative to the protons. Once the beam loading was complete, the beams were

accelerated to the top energy (in 86 seconds) and the machine optics was changed

to the collision configuration in 25 steps over 125 sec (low-beta squeeze). The

last two stages included initiating collisions at the two collision points and

removing halo by moving in the collimators. The experiments then commenced

data acquisition for the duration of the high-energy physics (HEP) store.

Figure 8: The collider fill cycle for store #8709 (May 2011).

Summary

For more than 25 years (1985-2011) the Fermilab antiproton complex was

the centerpiece of the Tevatron collider program [18] and provided antiprotons

for other particle physics experiments [19]. The continued Tevatron luminosity

increase was mainly due to a larger number of antiprotons being available, which

in turn was the result of a continuous and dedicated effort of hundreds of experts

to optimize and improve antiproton accumulation and cooling. The antiproton

stochastic and electron cooling methods were not invented at Fermilab, but they

were perfected to a degree not achieved anywhere else in the world.

12

References

[1] D. Cline et al., Proposal to construct an antiproton source for the Fermilab

accelerators, proposal 492, in Proceedings of 1976 NAL Summer Study on

Utilization of the Energy Doubler/Saver, Fermilab, Batavia U.S.A. (1976),

pg. 309.

[2] Tevatron Collider Run II Handbook, Ed. S.D. Holmes, Fermilab Preprint

TM-2484 (1998).

[3] C. D. Curtis et al., Linac H- Beam Operations and uses at Fermilab, IEEE

Transactions on Nuclear Science, NS-26, No. 3 (1979), 3760-3762.

[4] C.W. Schmidt, The Fermilab 400-MeV Linac Upgrade, in Proceedings of

1993 Particle Accelerator Conf., Washington, DC 1993, pp.1655–1659.

[5] E. Hubbard et al., Booster Synchrotron, Fermilab Technical Memo TM-405

(1973).

[6] S.D. Holmes, Design Criteria and Performance Goals for the Fermilab

Main Injector, Particle Accelerators, 58 (1997), 39.

[7] M. Church and J. Marriner, The Antiproton Sources: Design and Operation,

Annual Review of Nuclear and Particle Science, 43(1993), 253-295.

[8] R. J. Pasquinelli, Implementation of stochastic cooling hardware at

Fermilab’s Tevatron collider, JINST 6 T08002 (2011).

[9] D. McGinnis, Theory and design of microwave planar electrodes for

stochastic cooling of particle beams, Microw. Opt. Technol. Lett. 4 (1991)

439.

[10] D. McGinnis, Slotted waveguide slow-wave stochastic cooling arrays,

Proceedings of 1999 PAC, IEEE Conference, New York U.S.A. (1999), pg.

1713.

[11] V. Lebedev, Improvements to the stacktail and Debuncher momentum

cooling systems, in Proceedings of COOL 09, MOA1MCCO02, Lanzhou

China (2009).

[12] Fermilab Recycler Ring Technical Design Report, Ed. G. Jackson, Fermilab

Preprint TM-1991 (1997).

13

[13] S. Nagaitsev et al., Experimental Demonstration of Relativistic Electron

Cooling, Phys. Rev. Lett. 96, 044801 (2006).

[14] A. Burov et al., Optical principles of beam transport for relativistic electron

cooling, Phys. Rev. Spec. Top., Accel. Beams 3, 094002 (2000).

[15] A. Shemyakin and L.R. Prost, Ultimate performance of relativistic electron

cooling at Fermilab, in Proceedings of COOL11, THIOA01, Alushta,

Ukraine (2011).

[16] C.M. Bhat, Longitudinal momentum mining of beam particles in a storage

ring, Physics Letters A 330 (2004), p.481

[17] C. Gattuso et al., Optimization of Integrated Luminosity of the Tevatron, in

proceedings of PAC 2009, Vancouver, B. C., Canada.

[18] S. Holmes et al., Overview of the Tevatron collider complex: goals,

operations and performance, JINST 6 T08001 (2011).

[19] For example: Charmonium [E760/E835], Antihydrogen [E862], and APEX

[E868].